Bug 108976 - HDMI port detected as DVI on newer kernels
Summary: HDMI port detected as DVI on newer kernels
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: unspecified
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: Intel GFX Bugs mailing list
QA Contact: Intel GFX Bugs mailing list
Depends on:
Reported: 2018-12-08 08:22 UTC by zardam
Modified: 2019-06-07 18:07 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:

video bios (48.00 KB, application/octet-stream)
2018-12-08 08:22 UTC, zardam
no flags Details
dmesg on 4.9 kernel (82.11 KB, text/plain)
2019-01-09 16:39 UTC, zardam
no flags Details
dmesg on 4.18 kernel (98.32 KB, text/plain)
2019-01-09 16:39 UTC, zardam
no flags Details

Description zardam 2018-12-08 08:22:13 UTC
Created attachment 142749 [details]
video bios

I'm using debian stretch on a Shuttle SG33 system (Intel G33 Chipset).

With debian kernel 4.9.0-8-amd64, the two output ports are correctly discovered : one HDMI and one VGA: 

$ ls /sys/class/drm/
card0  card0-HDMI-A-1  card0-VGA-1  controlD64	renderD128  version

With debian kernel 4.18.0 (from stretch-backports), the HDMI port is discovered as DVI port, and the audio output no longer works on this port:

$ ls /sys/class/drm/
card0  card0-DVI-D-1  card0-VGA-1  renderD128  version

My "main" issue is the missing audio on the HDMI port when it is detected as DVI. The video is working properly.

I'm attaching the video bios, but I don't know it it is relevant, I can provide more information if needed.
Comment 1 Jani Nikula 2019-01-02 08:55:24 UTC
It would be helpful to get dmesgs on both kernels, with drm.debug=14 module parameter set, all the way from boot.
Comment 2 zardam 2019-01-09 16:39:10 UTC
Created attachment 143039 [details]
dmesg on 4.9 kernel
Comment 3 zardam 2019-01-09 16:39:30 UTC
Created attachment 143040 [details]
dmesg on 4.18 kernel
Comment 4 zardam 2019-01-09 16:40:33 UTC
I have attached the dmesg output on both kernel, with drm.debug=14
Comment 5 Ville Syrjala 2019-01-17 19:10:05 UTC
This was broken by de44e256b92c ("drm/i915/sdvo: Shut up state checker with hdmi cards on gen3")

I have a branch which should hopefully restore things to a working state (and improve a bunch of other things as well):
git://github.com/vsyrjala/linux.git sdvo_limited_range_5

I'd quite like a gen3 machine with SDVO HDMI for testing, but after checking ebay no way. 150-250 bucks for a g33. Nuts!
Comment 6 zardam 2019-01-19 22:46:08 UTC
Thank you for your answer.

I tested your branch, and the output is now detected again as HDMI in /sys/class/drm/

I cannot test audio output as the kernel does not want to mount the root filesystem (over iscsi) for now.
Comment 7 zardam 2019-01-20 18:48:21 UTC
I confirm, audio over HDMI is working fine with you branch (I finally managed to boot the system).
Comment 8 Ville Syrjala 2019-06-07 18:07:18 UTC
Should now be fixed:

commit dc49a56bd43bb04982e64b44436831da801d0237
Author: Ville Syrjälä <ville.syrjala@linux.intel.com>
Date:   Tue Apr 9 17:40:49 2019 +0300

    drm/i915/sdvo: Implement proper HDMI audio support for SDVO

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.